Please Tell the Writer What He or She Did Well and What He or She Needs to Work On: A Feedback Science Explainer

Please tell the writer what he or she did well and what he or she needs to work on in specific, behavioral terms, because vague praise and vague criticism both fail to change the next draft; effective feedback names the observable behavior, ties it to a concrete effect on the reader, and proposes one actionable next step.

Feedback works when it is specific, timely, and aimed at behavior rather than identity. A comment such as "your transitions are abrupt" gives the writer something to fix. A comment such as "you are a confusing writer" gives the writer something to defend against. The difference is not politeness for its own sake. It is the difference between information a person can use and information a person must emotionally manage first. This explainer covers the mechanism behind that difference, the research context, common edge cases, misconceptions, practical steps, and the limits of what feedback science can promise.

Mechanism: Why Specific Feedback Changes Behavior

Feedback improves performance through a loop with three links: attention, comparison, and adjustment. The writer must first notice the gap between intended and actual effect. Then they must compare that gap against a standard they accept. Finally they must have a feasible next move. Break any link and the feedback produces anxiety or defensiveness instead of revision.

The neural and behavioral basis is error-driven learning. When a person receives information that conflicts with their expectation, the brain allocates attention to reconcile the mismatch (Hattie & Timperley, 2007). Vague feedback creates no mismatch signal because there is nothing concrete to compare. "Make it stronger" cannot be reconciled with anything. "Your second paragraph buries the main claim under two sentences of setup" creates a precise mismatch the writer can act on.

Feedback also works through goal orientation. Feedback that targets the task ("this opening answers the question in sentence one") supports mastery goals. Feedback that targets the person ("you are a natural") shifts attention to protecting a reputation. The person-targeted version feels better and works worse. It teaches writers that the point of writing is to be seen as talented rather than to make a reader understand something.

One counterintuitive detail matters here. Praise is not the opposite of criticism; it is a separate channel. A draft can need both, but they should be sequenced so the writer knows which comments are diagnostic and which are encouragement. Mixing them into one sentence ("great energy, but this section is unclear") often leaves the writer remembering only the compliment.

Quantified Context: What the Research Shows

The most cited framework in feedback research distinguishes three questions every feedback message should answer: where am I going, how am I going, and where to next (Hattie & Timperley, 2007). Feedback that answers none of these three produces little measurable effect. Feedback that answers the first and second without the third leaves writers knowing they missed but not how to correct course.

Effect sizes in education research cluster around feedback as one of the highest-leverage interventions, but the range is wide, roughly from near zero to well above average depending on type. The pattern is consistent: task-level and process-level feedback outperform praise, punishment, and self-level comments. This is why a review that says "strong work" is thin even when sincere.

Timing interacts with type. Immediate feedback helps procedural tasks where the correct move is clear. Delayed feedback helps complex writing tasks because the writer needs distance to see the text as a reader would. For a 2,000-word draft, comments delivered within a day or two of submission tend to land better than comments delivered weeks later after the writer has mentally moved on.

Volume has diminishing returns. A reviewer who leaves 60 comments on a short draft often produces paralysis rather than revision. Prioritizing the 5 to 8 comments that address structure and clarity usually produces more revision than exhaustive line editing.

Timeline: How a Feedback Cycle Should Run

The timeline matters because feedback is not a document; it is a conversation with a rhythm. Reviewers who dump comments and disappear leave writers guessing which notes are essential. Writers who revise everything at once lose track of which change addressed which problem.

Edge Cases: When Standard Advice Fails

Several situations break the usual rules, and a good reviewer adjusts rather than recites the framework. A writer under acute stress or facing a deadline may need one or two priority notes, not a full critique. A draft that is already strong may need only a note about what to protect in revision. A writer learning a new genre may need models and examples before critique.

Cultural and power dynamics also change the calculus. Feedback from a supervisor carries implicit stakes that feedback from a peer does not, and comments that read as neutral in one context can read as permission-withholding in another. Reviewers with authority should separate "this is wrong" from "this is not my preference" explicitly, because subordinates often cannot ask.

Neurodivergent writers, writers working in a second language, and writers with anxiety around evaluation may need feedback in writing rather than in a meeting, or in smaller batches. None of this changes the core principles. It changes the delivery.

Red-flag warnings for reviewers: Comments that describe the reader's reaction without pointing to text ("this felt rambling") are hard to act on unless you cite the passage. Sarcasm and rhetorical questions ("did you even read this?") shut down revision. Editing commas while the structure is broken wastes the writer's attention. Praising only what is easy to praise teaches writers to optimize for the wrong things. Feedback delivered in public creates performance pressure that suppresses honest questions. A review with no stated priority leaves the writer to guess what matters most.

Misconceptions: What Feedback Is Not

The most common misconception is that feedback means evaluation. Evaluation ranks a draft; feedback changes the next one. A reviewer can do both, but conflating them means the writer hears a verdict when they needed a direction. A grade with no behavioral notes tells the writer where they landed, not how to move.

A second misconception is that the "sandwich" (praise, criticism, praise) is required. The evidence does not support a fixed ratio, and forced praise can read as manipulative to experienced writers. What matters is clarity about which comments are diagnostic. A short note that says "these three comments are the ones to act on first" does more than any sandwich.

A third misconception is that writers want only positive feedback. Writers generally want to improve, and they can handle directness when it is specific and respectful. What they cannot handle is ambiguity, because ambiguity forces them to guess what the reviewer actually meant.

A fourth misconception is that feedback should be exhaustive. Thoroughness is a virtue in proofreading and a vice in structural review. The reviewer's job is to find the few changes that would most improve the draft, not to demonstrate that they read every word.

Practical Actions: A Concrete Protocol

Use a short protocol and adapt it. The goal is not a perfect system but a repeatable one that produces specific, prioritized, actionable notes.

Before reading, ask the writer what they were trying to do and what kind of feedback they want. This single question prevents most mismatches. Read the whole draft once without commenting. Note your reactions, but do not write them down yet. On the second pass, mark passages where you lost the thread, where a claim went unsupported, and where the structure stopped making sense. These are the high-value notes. Write each comment in three parts: the observable behavior, the effect on you as a reader, and one suggested next step. "The second paragraph opens with two sentences of setup before the claim; I did not know what the section was about until sentence three; consider leading with the claim." That is a usable note. Rank the comments. Mark the top three to five as priority. Everything else is optional polish. Close with one question the writer should answer in revision. A question ("what do you want the reader to do after the last paragraph?") often does more than a directive. If you are the writer receiving feedback, read all comments before responding. Separate comments about structure from comments about style. Revise one category at a time.

For reviewers who want a threshold: if a draft has more than roughly 20 substantive comments, you are probably doing the writer's job rather than giving feedback. Aim for fewer, deeper notes.

For writers requesting feedback, name the stage of the draft ("this is a first draft, I want structural notes, not line edits") and name the deadline you actually have. Specific requests produce specific feedback.

Reconciling thresholds across sources is straightforward. Educational research emphasizes task and process feedback over praise. Professional writing practice emphasizes prioritization and specificity. Both point to the same behavior: fewer comments, aimed at the highest-leverage issues, each with a concrete next step. Where sources differ is on timing and volume, and those depend on the task, the relationship, and the stakes.

Limitations: What Feedback Science Cannot Tell You

Feedback research is strong on averages and weak on individuals. An effect size tells you what tends to work across many writers and tasks; it does not tell you what will work for this writer on this draft. Treat the principles as defaults, not rules.

Much of the strongest evidence comes from educational settings with structured tasks, which differ from open-ended creative or professional writing. The mechanisms likely transfer, but the specific thresholds (how many comments, how soon) are less settled. Any number in this article should be read as a working range, not a measured constant.

Feedback also cannot substitute for the writer's own judgment. A reviewer can point to a problem, but the writer decides whether the note is right, which change to make, and how to preserve what is working. Good feedback increases the writer's options; it does not remove the writer's authority over the text.

No study can tell you whether a particular comment will land as helpful or as harsh, because that depends on the relationship, the context, and the writer's state on the day. The best a reviewer can do is be specific, be prioritized, be respectful, and be available for questions.

This article is general educational information about communication and learning, not a clinical, therapeutic, or professional evaluation of any individual writer or reviewer. It does not diagnose, treat, or promise improved performance or outcomes. For concerns about learning difficulties, anxiety around evaluation, or workplace communication problems, consult a qualified local professional.